A python script that generates a report about contributions to a Git repository.

git-contribution.py

#!/usr/bin/env python

"""Git contributions

Reference: https://gist.github.com/ajardin/b9e0d33ec8c3a3b2874d

The script below allows to have a quick overview of contributions on a Git repository.
Especially useful when we do not have access to graphs like those that we can find on Github.

This script handles three non-mandatory parameters:
   - START is used to keep commits only since the given date (default = 01 Jan 2000).
   - END is used to keep commits only until the given date (default = 01 Jan 2050).
   - MINIMUM is used to define the minimum number of commits in order to appear in the report (default = 10).
"""

import optparse
import sys
import os
import re
import operator
import unicodedata

from subprocess import Popen, PIPE, check_output

__version__ = "0.1"

PATH_TO_GIT = "." # CHANGE IT FOR YOUR GIT PROJECT LOCATION
USAGE = "%prog [options]"
VERSION = "%prog v" + __version__


def strip_accents(text):
    """
    Strip accents from input String.

    :param text: The input string.
    :type text: String.

    :returns: The processed String.
    :rtype: String.
    """
    try:
        text = unicode(text, 'utf-8')
    except NameError: # unicode is a default on python 3 
        pass
    text = unicodedata.normalize('NFD', text)
    text = text.encode('ascii', 'ignore')
    text = text.decode("utf-8")
    return str(text)


def remove_accents(input_str):
    nkfd_form = unicodedata.normalize('NFKD', input_str)
    return u"".join([c for c in nkfd_form if not unicodedata.combining(c)])


def extract_stats(stats, minimum):
    p = re.compile(ur'^\s*(\d+)\s+(.+)\s+<(.+)>$', re.MULTILINE)
    raw_stats = {}
    result = []

    for stat in stats:
        res = re.search(p, stat)
        if res:
            author_commits = int(res.group(1))
            author_name = strip_accents(res.group(2))
            author_mail = res.group(3)

            if not author_mail in raw_stats.keys():
                raw_stats[author_mail] = {'name': author_name, 'commits': 0}

            raw_stats[author_mail]['commits'] += author_commits

    sorted_x = list(reversed(sorted(raw_stats.items(), key=operator.itemgetter(1))))
    for val in sorted_x:
        if val[1]['commits'] >= minimum:
            result.append((val[0], val[1]['name'], val[1]['commits']))

    return result


def format_report(stats, since, before):
    output = "\n%-5s\t%-20s\t%-33s\t%-8s\t%-10s\t%-10s\t" % ("Rank", "Name", "Email", "Commits", "Files Changed", "Insertions (+)")
    output += "%-10s\t%-18s\t%-24s\t%-24s" % ("Deletions (-)", "Diff/Commits (~)", "First Commit", "Last Commit")
    print output

    rank=0
    for stat in stats:
        rank += 1

        author_mail = stat[0]
        author_name = stat[1]
        author_commits = stat[2]
        files = 0
        insertions = 0
        deletions = 0

        cmd_data = 'git log --shortstat --author="%s" --since="%s" --before="%s" --no-merges' % (author_mail, since, before)
        cmd_data += ' | grep -E "fil(e|es) changed"'
        cmd_data += ' | awk \'{ files+=$1; inserted+=$4; deleted+=$6 } END { print "FILES="files, "INSERTIONS="inserted, "DELETIONS="deleted }\''

        tmp = check_output(cmd_data, shell=True).rstrip('\n').split(" ")
        data = {
            'files': int(tmp[0].split("=")[1]),
            'insertions': int(tmp[1].split("=")[1]),
            'deletions': int(tmp[2].split("=")[1])
        }
        data['avg_commits'] = (data['insertions'] + data['deletions']) / author_commits

        cmd_first_commit_hash = 'git log --author="%s" --since="%s" --before="%s" --no-merges --reverse | head -1' % (author_mail, since, before)
        first_commit_hash = check_output(cmd_first_commit_hash, shell=True).rstrip('\n').split(" ")[1]

        cmd_first_commit_date = 'git show -s --format=%%ar "%s"' % (first_commit_hash)
        first_commit_date = check_output(cmd_first_commit_date, shell=True).rstrip('\n')

        cmd_last_commit_hash = 'git log --author="%s" --since="%s" --before="%s" --no-merges | head -1' % (author_mail, since, before)
        last_commit_hash = check_output(cmd_last_commit_hash, shell=True).rstrip('\n').split(" ")[1]

        cmd_last_commit_date = 'git show -s --format=%%ar "%s"' % (last_commit_hash)
        last_commit_date = check_output(cmd_last_commit_date, shell=True).rstrip('\n')

        output = "%-5s\t%-20s\t%-33s\t%-8s\t%-10s\t%-10s\t%-10s" % (rank, author_name, author_mail, author_commits, data['files'], data['insertions'], data['deletions'])
        output += "\t%-18s\t%-24s\t%-24s" % (data['avg_commits'], first_commit_date, last_commit_date)

        print(output)

    print("")


def parse_options():
    """parse_options() -> opts, args

    Parse any command-line options given returning both
    the parsed options and arguments.
    """

    parser = optparse.OptionParser(usage=USAGE, version=VERSION)

    parser.add_option("-s", "--start",
            action="store", type="string", default="01 Jan 2000", dest="start",
            help="Keep commits only since the given date (default = 01 Jan 2000)")

    parser.add_option("-e", "--end",
            action="store", type="string", default="01 Jan 2050", dest="end",
            help="Keep commits only until the given date (default = 01 Jan 2050)")

    parser.add_option("-m", "--minimum",
            action="store", type="int", default=10, dest="minimum",
            help="Define the minimum number of commits in order to appear in the report")

    parser.add_option("-v", "--verbose", action="store_true", default=False,
                      dest="verbose", help="Launch in verbose mode")

    opts, args = parser.parse_args()

    return opts, args


def main():    
    opts, args = parse_options()

    os.chdir(PATH_TO_GIT)
    process = Popen(['git', 'shortlog', '-sne', '--since="%s"' % (opts.start), '--before="%s"' % (opts.end), '--no-merges'], stdout=PIPE, stderr=PIPE)
    stdout, stderr = process.communicate()

    if len(stdout):
        stats = extract_stats(stdout.split('\n'), opts.minimum)
        format_report(stats, opts.start, opts.end)


if __name__ == "__main__":
    main()